This document proposes a method for optimally placing a jammer node in a wireless localization network to degrade the accuracy of locating target nodes. It formulates the problem of maximizing the minimum Cramér-Rao lower bound (CRLB) for target nodes as an optimization with constraints on jammer location. Theoretical results show the optimal jammer placement is determined by one, two or three target nodes depending on scenarios. Explicit expressions are derived for optimal placement with two targets. Numerical examples illustrate key results.
Low rank tensor approximation of probability density and characteristic funct...Alexander Litvinenko
Very often one has to deal with high-dimensional random variables (RVs). A high-dimensional RV can be described by its probability density (\pdf) and/or by the corresponding probability characteristic functions (\pcf), or by a function representation. Here the interest is mainly to compute characterisations like the entropy, or
relations between two distributions, like their Kullback-Leibler divergence, or more general measures such as $f$-divergences,
among others. These are all computed from the \pdf, which is often not available directly, and it is a computational challenge to even represent it in a numerically feasible fashion in case the dimension $d$ is even moderately large. It is an even stronger numerical challenge to then actually compute said characterisations in the high-dimensional case.
In this regard, in order to achieve a computationally feasible task, we propose to represent the density by a high order tensor product, and approximate this in a low-rank format.
Low-rank matrix approximations in Python by Christian Thurau PyData 2014PyData
Low-rank approximations of data matrices have become an important tool in machine learning and data mining. They allow for embedding high dimensional data in lower dimensional spaces and can therefore mitigate effects due to noise, uncover latent relations, or facilitate further processing. These properties have been proven successful in many application areas such as bio-informatics, computer vision, text processing, recommender systems, social network analysis, among others. Present day technologies are characterized by exponentially growing amounts of data. Recent advances in sensor technology, internet applications, and communication networks call for methods that scale to very large and/or growing data matrices. In this talk, we will describe how to efficiently analyze data by means of matrix factorization using the Python Matrix Factorization Toolbox (PyMF) and HDF5. We will briefly cover common methods such as k-means clustering, PCA, or Archetypal Analysis which can be easily cast as a matrix decomposition, and explain their usefulness for everyday data analysis tasks.
Hierarchical matrix techniques for maximum likelihood covariance estimationAlexander Litvinenko
1. We apply hierarchical matrix techniques (HLIB, hlibpro) to approximate huge covariance matrices. We are able to work with 250K-350K non-regular grid nodes.
2. We maximize a non-linear, non-convex Gaussian log-likelihood function to identify hyper-parameters of covariance.
IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology
Low rank tensor approximation of probability density and characteristic funct...Alexander Litvinenko
Very often one has to deal with high-dimensional random variables (RVs). A high-dimensional RV can be described by its probability density (\pdf) and/or by the corresponding probability characteristic functions (\pcf), or by a function representation. Here the interest is mainly to compute characterisations like the entropy, or
relations between two distributions, like their Kullback-Leibler divergence, or more general measures such as $f$-divergences,
among others. These are all computed from the \pdf, which is often not available directly, and it is a computational challenge to even represent it in a numerically feasible fashion in case the dimension $d$ is even moderately large. It is an even stronger numerical challenge to then actually compute said characterisations in the high-dimensional case.
In this regard, in order to achieve a computationally feasible task, we propose to represent the density by a high order tensor product, and approximate this in a low-rank format.
Low-rank matrix approximations in Python by Christian Thurau PyData 2014PyData
Low-rank approximations of data matrices have become an important tool in machine learning and data mining. They allow for embedding high dimensional data in lower dimensional spaces and can therefore mitigate effects due to noise, uncover latent relations, or facilitate further processing. These properties have been proven successful in many application areas such as bio-informatics, computer vision, text processing, recommender systems, social network analysis, among others. Present day technologies are characterized by exponentially growing amounts of data. Recent advances in sensor technology, internet applications, and communication networks call for methods that scale to very large and/or growing data matrices. In this talk, we will describe how to efficiently analyze data by means of matrix factorization using the Python Matrix Factorization Toolbox (PyMF) and HDF5. We will briefly cover common methods such as k-means clustering, PCA, or Archetypal Analysis which can be easily cast as a matrix decomposition, and explain their usefulness for everyday data analysis tasks.
Hierarchical matrix techniques for maximum likelihood covariance estimationAlexander Litvinenko
1. We apply hierarchical matrix techniques (HLIB, hlibpro) to approximate huge covariance matrices. We are able to work with 250K-350K non-regular grid nodes.
2. We maximize a non-linear, non-convex Gaussian log-likelihood function to identify hyper-parameters of covariance.
IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology
Overlap Add, Overlap Save(digital signal processing)Gourab Ghosh
In DSP to solve a convolution of a long duration sequence there are two popular methods. Overlap Add, Overlap Save. In this presentation i've discussed about both.
- Gourab Ghosh
Identification of unknown parameters and prediction of missing values. Compar...Alexander Litvinenko
H-matrix approximation of large Mat\'{e}rn covariance matrices, Gaussian log-likelihoods.
Identifying unknown parameters and making predictions
Comparison with machine learning methods.
kNN is easy to implement and shows promising results.
Application of Module Structure of Algebra in Homomorphic Signal Processingijsrd.com
In this paper we see how the module structure of Algebra plays a vital role in Homomorphic Signal Processing in branch of Engineering mainly in Information Technology, Electronics and Telecommunication Engineering, Computer Science etc. Here I generalize this concept for module theory in algebra and I introduce the concept module logic. I give some results on module logic and also give conclusion that why the module logic is useful in Signal Processing.
Positive and negative solutions of a boundary value problem for a fractional ...journal ijrtem
: In this work, we study a boundary value problem for a fractional
q, -difference equation. By
using the monotone iterative technique and lower-upper solution method, we get the existence of positive or
negative solutions under the nonlinear term is local continuity and local monotonicity. The results show that we
can construct two iterative sequences for approximating the solutions
Overlap Add, Overlap Save(digital signal processing)Gourab Ghosh
In DSP to solve a convolution of a long duration sequence there are two popular methods. Overlap Add, Overlap Save. In this presentation i've discussed about both.
- Gourab Ghosh
Identification of unknown parameters and prediction of missing values. Compar...Alexander Litvinenko
H-matrix approximation of large Mat\'{e}rn covariance matrices, Gaussian log-likelihoods.
Identifying unknown parameters and making predictions
Comparison with machine learning methods.
kNN is easy to implement and shows promising results.
Application of Module Structure of Algebra in Homomorphic Signal Processingijsrd.com
In this paper we see how the module structure of Algebra plays a vital role in Homomorphic Signal Processing in branch of Engineering mainly in Information Technology, Electronics and Telecommunication Engineering, Computer Science etc. Here I generalize this concept for module theory in algebra and I introduce the concept module logic. I give some results on module logic and also give conclusion that why the module logic is useful in Signal Processing.
Positive and negative solutions of a boundary value problem for a fractional ...journal ijrtem
: In this work, we study a boundary value problem for a fractional
q, -difference equation. By
using the monotone iterative technique and lower-upper solution method, we get the existence of positive or
negative solutions under the nonlinear term is local continuity and local monotonicity. The results show that we
can construct two iterative sequences for approximating the solutions
A Mathematically Derived Number of Resamplings for Noisy Optimization (GECCO2...Jialin LIU
"A Mathematically Derived Number of Resamplings for Noisy Optimization". Jialin Liu, David L. St-Pierre and Olivier Teytaud. (Accepted as short paper) Genetic and Evolutionary Computation Conference (GECCO), 2014.
Response Surface in Tensor Train format for Uncertainty QuantificationAlexander Litvinenko
We apply low-rank Tensor Train format to solve PDEs with uncertain coefficients. First, we approximate uncertain permeability coefficient in TT format, then the operator and then apply iterations to solve stochastic Galerkin system.
Joint blind calibration and time-delay estimation for multiband rangingTarik Kazaz
In this presentation, we focus on the problem of blind joint calibration of multiband transceivers and time-delay (TD) estimation of multipath channels. We show that this problem can be formulated as a particular case of covariance matching. Although this problem is severely ill-posed, prior information about radio-frequency chain distortions and multipath channel sparsity is used for regularization. This approach leads to a biconvex optimization problem, which is formulated as a rank-constrained linear system and solved by a simple group Lasso algorithm.
% This method is general and can be also applied for calibration of sensors arrays and in direction of arrival estimation.
Numerical experiments show that the proposed algorithm provides better calibration and higher resolution for TD estimation than current state-of-the-art methods.
Talk on Resource Allocation Strategies for Layered Multimedia Multicast ServicesAndrea Tassi
The explosive growth of content-on-the-move, such as video streaming to mobile devices, has propelled research on multimedia broadcast and multicast schemes. Multi-rate transmission strategies have been proposed as a means of delivering layered services to users experiencing different downlink channel conditions. In this presentation, we consider random linear network coding for its inherent reliability features and study two encoding approaches, which are appropriate for layered services. We derive packet error probability expressions and use them as performance metrics in the formulation of resource allocation frameworks. The aim of these frameworks is both the optimization of the transmission scheme and the minimization of the number of broadcast packets on each downlink channel, while offering service guarantees to a predetermined fraction of users. Our proposed frameworks are adapted to the LTE stack and the integrated eMBMS technology. We focus on the delivery of a video service based on the H.264/SVC standard and demonstrate the advantages of layered network coding over multi-rate transmission. Furthermore, we establish that the choice of both the network coding technique and the resource allocation method play a critical role in the footprint of a service, as determined by the quality of each received video layer.
Subgradient Methods for Huge-Scale Optimization Problems - Юрий Нестеров, Cat...Yandex
We consider a new class of huge-scale problems, the problems with sparse subgradients. The most important functions of this type are piecewise linear. For optimization problems with uniform sparsity of corresponding linear operators, we suggest a very efficient implementation of subgradient iterations, the total cost of which depends logarithmically in the dimension. This technique is based on a recursive update of the results of matrix/vector products and the values of symmetric functions. It works well, for example, for matrices with few nonzero diagonals and for max-type functions.
We show that the updating technique can be efficiently coupled with the simplest subgradient methods. Similar results can be obtained for a new non-smooth random variant of a coordinate descent scheme. We also present promising results of preliminary computational experiments.
Possible applications of low-rank tensors in statistics and UQ (my talk in Bo...Alexander Litvinenko
Just some ideas how low-rank matrices/tensors can be useful in spatial and environmental statistics, where one usually has to deal with very large data
Computation of electromagnetic fields scattered from dielectric objects of un...Alexander Litvinenko
We develop fast and efficient stochastic methods for characterizing scattering
from objects of uncertain shapes. This is highly needed in the
fields of electromagnetics, optics, and photonics.
The continuation multilevel Monte Carlo (CMLMC) method is
used together with a surface integral equation solver. The
CMLMC method optimally balances statistical errors due to
sampling of the parametric space, and numerical errors due
to the discretization of the geometry using a hierarchy of
discretizations, from coarse to fine. The number of realizations
of finer discretizations can be kept low, with most samples
computed on coarser discretizations to minimize computational
work. Consequently, the total execution time is significantly
reduced, in comparison to the standard MC scheme.
MVPA with SpaceNet: sparse structured priorsElvis DOHMATOB
The GraphNet (aka S-Lasso), as well as other “sparsity + structure” priors like TV (Total-Variation), TV-L1, etc., are not easily applicable to brain data because of technical problems
relating to the selection of the regularization parameters. Also, in
their own right, such models lead to challenging high-dimensional optimization problems. In this manuscript, we present some heuristics for speeding up the overall optimization process: (a) Early-stopping, whereby one halts the optimization process when the test score (performance on leftout data) for the internal cross-validation for model-selection stops improving, and (b) univariate feature-screening, whereby irrelevant (non-predictive) voxels are detected and eliminated before the optimization problem is entered, thus reducing the size of the problem. Empirical results with GraphNet on real MRI (Magnetic Resonance Imaging) datasets indicate that these heuristics are a win-win strategy, as they add speed without sacrificing the quality of the predictions. We expect the proposed heuristics to work on other models like TV-L1, etc.
1. Optimal Jammer Placement in Wireless Localization Networks
Sinan Gezici∗
, Suat Bayram♭
, Mohammad Reza Gholami♮
, and Magnus Jansson♮
∗ Department of Electrical and Electronics Engineering, Bilkent University, 06800, Ankara, Turkey
♭ Department of Electrical and Electronics Engineering, Turgut Ozal University, 06010, Ankara, Turkey
♮ ACCESS Linnaeus Centre, Department of Signal Processing, KTH–Royal Institute of Technology, Stockholm, Sweden
Abstract
The optimal jammer placement problem is proposed for a wireless localization network, where
the aim is to degrade the accuracy of locating target nodes as much as possible. In particular, the
optimal location of a jammer node is obtained in order to maximize the minimum of the Cram´er-
Rao lower bounds for a number of target nodes under location related constraints for the jammer
node. Theoretical results are derived to specify scenarios in which the jammer node should be
located as close to a certain target node as possible, or the optimal location of the jammer node is
determined by two or three of the target nodes. In addition, explicit expressions for the optimal
location of the jammer node are derived in the presence of two target nodes. Numerical examples
are presented to illustrate the theoretical results.
Introduction
• Position information: vital for many location aware services/applications
• The main aim in wireless localization networks: achieving high localization accuracy
• Wireless sensor localization: a well investigated topic
• The effects of jamming on wireless localization networks: a rather new topic - little
attention in the literature
• Recent work [1]: the optimal power allocation strategies for the jammer nodes in or-
der to maximize the average or the minimum Cramr-Rao lower bounds (CRLBs) of
the target nodes.
• Current work: optimal jammer placement problem based on maximization the mini-
mum CRLB
System Model
• NA anchor nodes at known locations yi ∈ R2
and NT target nodes at xi ∈ R2
• A jammer node at z ∈ R2
• The jammer node is assumed to transmit zero-mean Gaussian noise
• Ai {j ∈ {1,...,NA} | anchor node j is connected to target node i}, i ∈ {1,...,NT}
Received signal: ri j(t) =
Lij
∑
k=1
αk
i js(t −τk
i j)+γi
√
PJ vi(t)+ni j(t), t ∈ [0,Tobs], j ∈ Ai
ni j(t),
√
PJ vi(t): independent zero-mean white Gaussian random processes
• τk
i j
yj−xi +bk
ij
c bk
i j ≥ 0 & Ai A L
i ∪A NL
i
• bi j [b2
i j ...b
Lij
i j ] if j ∈ A L
i & bi j [b1
i j ...b
Lij
i j ]T
if j ∈ A NL
i
• θi [xT
i bT
iAi(1) ··· bT
iAi(|Ai|)]T
• CRLB: E{ ˆxi −xi
2
} ≥ tr{[F−1
i ]2×2} with [F−1
i ]2×2 = Ji(xi,PJ)−1
• Ref.[2] : Ji(xi,PJ) = ∑
j∈A L
i
λij
N0/2+PJ|γi|2
φijφT
ij λi j
4π2
β2
|α1
ij|2 ∞
−∞ |S( f)|2
d f
c2 (1−ξj)
φi j [cosϕi j sinϕi j]T
β =
∞
−∞ f2|S( f)|2d f
∞
−∞ |S( f)|2d f 0 ≤ ξj ≤ 1: the path-overlap coefficient
ϕi j: the angle between target node i and anchor node j
• CRLBi = tr{Ji(xi,PJ)−1
} = ri(PJ|γi|2
+N0/2) ri tr{[∑j∈A L
i
λi jφi jφT
i j]−1
}
Optimal Jammer Placement
Generic Formulation and Analysis
maximize
z
min
i∈{1,...,NT }
ri PJ|γi|2
+
N0
2
subject to z−xi ≥ ε , i = 1,...,NT
Assumption: |γi|2
= ˜Ki
d0
z−xi
ν
for z−xi > d0
maximize
z
min
i∈{1,...,NT }
ri
KiPJ
z−xi
ν
+
N0
2
subject to z−xi ≥ ε , i = 1,...,NT
(1)
Ki ˜Ki(d0)ν
Proposition 1: If there exists a target node that satisfies the following inequality,
rℓ
KℓPJ
εν
+
N0
2
≤ min
i∈{1,...,NT }
i=ℓ
ri
KiPJ
( xi −xℓ +ε)ν
+
N0
2
(2)
and if set {z : z−xℓ = ε & z−xi ≥ ε, i = 1,...,ℓ−1,ℓ+1,...,NT} is
non-empty, then the solution of (1), denoted by zopt
, satisfies zopt
−xℓ = ε; that is,
the jammer node is placed at a distance of ε from the ℓth target node.
The optimization problem in (1) in the presence of two target nodes ℓ1 and ℓ2 only:
maximize
z
min
i∈{ℓ1,ℓ2}
ri
KiPJ
z−xi
ν
+
N0
2
subject to z−xℓ1
≥ ε , z−xℓ2
≥ ε
(3)
ℓ1,ℓ2 ∈ {1,...,NT} and ℓ1 = ℓ2. zopt
ℓ1,ℓ2
and CRLBℓ1,ℓ2
: the optimizer and the optimal
value of (3)
Proposition 2: Let CRLBk,i be the minimum of CRLBℓ1,ℓ2
for ℓ1,ℓ2 ∈ {1,...,NT} and
ℓ1 = ℓ2, and let zopt
k,i denote the corresponding jammer location (i.e., the optimizer of
(3) for ℓ1 = k and ℓ2 = i). Then, an optimal jammer location obtained from (1) is
equal to zopt
k,i if zopt
k,i is an element of set z : z−xm ≥ ε, m ∈ {1,...,NT}{k,i} and
rm
KmPJ
zopt
k,i −xm
ν
+
N0
2
≥ CRLBk,i (4)
for m ∈ {1,...,NT}{k,i}.
⇒ the optimal jammer location is mainly determined by two of the target nodes since
the others have larger CRLBs when the jammer node is placed at the optimal location
according to those two jammer nodes only.
——————————–Special cases———————————–
1-Two Target Nodes
Proposition 3: For the case of two target nodes (i.e., NT = 2), the solution zopt
of (1)
satisfies one of the following conditions:
(i) if x1 −x2 < 2ε, then zopt
−x1 = zopt
−x2 = ε
(ii) otherwise,
(a) if r1
K1PJ
εν + N0
2 ≤ r2
K2PJ
( x1−x2 −ε)ν + N0
2 , then zopt
−x1 = ε and
zopt
−x2 = x1 −x2 −ε
(b) if r2
K2PJ
εν + N0
2 ≤ r1
K1PJ
( x1−x2 −ε)ν + N0
2 , then zopt
−x1 = x1 −x2 −ε and
zopt
−x2 = ε
(c) otherwise, zopt
−x1 = d∗
and zopt
−x2 = x1 −x2 −d∗
, where d∗
is the unique
solution of the following equation over d ∈ (ε, x1 −x2 −ε).
r1
K1PJ
dν
+
N0
2
= r2
K2PJ
( x1 −x2 −d)ν
+
N0
2
(5)
2-Infinitesimally Small ε
Proposition 4: Suppose NT ≥ 3 and ε → 0. The max-min CRLB in the presence of
target nodes ℓ1, ℓ2, and ℓ3: CRLBℓ1,ℓ2,ℓ3
= maxz minm∈{ℓ1,ℓ2,ℓ3} rm
KmPJ
z−xm
ν + N0
2 . Also,
let target nodes i, j, and k achieve the minimum of CRLBℓ1,ℓ2,ℓ3
for
ℓ1,ℓ2,ℓ3 ∈ {1,...,NT } and let zopt
i, j,k denote the jammer location corresponding to
CRLBi,j,k. Then, the optimal location for the jammer node (i.e., the optimizer of (1) in
the absence of the distance constraints) is equal to zopt
i, j,k, and at least two of the CRLBs
of the target nodes are equalized by the optimal solution.
Numerical Examples
• Simulation parameters: ε = 1m, PJ = 6, N0 = 2, ν = 2, and Ki = 1 for i = 1,...,NT.
LOS scenarios, λi j = 100N0 xi −yj
−2
/2 (free space propagation model)
0 1 2 3 4 5 6 7 8 9 10
0
1
2
3
4
5
6
7
8
9
10
Optimal location of the jammer node
x [m]
y[m]
Anchor node
Target node
Target 1
Target 2
Target 3
Target 4
ε
ε
ε
ε
0 1 2 3 4 5 6 7 8 9 10
0
1
2
3
4
5
6
7
8
9
10
Optimal location of the jammer node
x [m]
y[m]
Anchor node
Target node
Target 1
Target 2
Target 3
Target 4
ε
ε
ε
ε
Fig. 1. The network consisting of anchor nodes at [0 0], [10 0], [0 10], and [10 10] m. Left figure: target nodes at [2 6], [5 2], [8 3], and [9 6] m.
Right figure target nodes at [2 5], [4 1], [8 8], and [9 2] m.
V. N E• Left Fig: Proposition 2 for k = 1&i = 3 → the solution determined by subnetwork
consisting of target 1 and 3. Solution: zopt
1,3 = [5.0605 4.4697] m, CRLB1,3 = 0.8053m2
• Right Fig: Subnetwork consisting of taget nodes 1,3, and 4 archives the minimum
max min CRLB among all other possible subnetworks with three target nodes.
Solution: CRLB1,3,4 = 0.7983m2
and zopt
1,3,4 = [5.5115 5.5717]m
References
[1] S. Gezici, M. R. Gholami, S. Bayram, and M. Jansson, “Optimal jamming of wireless localization
systems,” in IEEE International Conference on Communications (ICC) Workshops, June 2015.
[2] Y. Shen, and MZ. Win, “Fundamental limits of wideband localizationPart I: A general framework,”
IEEE Transactions on Information Theory, pp. 4956-4980, 2010.
Emails: gezici@ee.bilkent.edu.tr, sbayram@turgutozal.edu.tr, mohrg@kth.se, janssonm@kth.se